Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
Time of Flight Mass Spectrometry of Ions Generated
by Molecules in Intense Laser Fields
Mingtong Han
25th
August, 2013
Abstract
Photoionization of acetylene, dimethylacetylene and methylacetylene in an intense laser field (45
fs, 800 nm, 15 210 W/cm ) is studied by high-resolution time of flight mass spectrometry (TOFMS).
Possible pathways of Coulomb explosions during photoionization of acetylene are studied by careful
analysis of the signal profiles and momentum releases of the fragments. The laser field intensity
dependence and polarization dependence of Coulomb explosion fragments of acetylene are investigated
by mass-resolved momentum imaging (MRMI) of the ion fragments under a sequence of different laser
field intensities and rotated laser polarization directions. New features of ejection of triatomic hydrogen
ions in photoionization of dimethylacetylene and methylacetylene molecules are also observed and
recorded.
1. Introduction
Ultrafast laser highly increases the intensity of laser field that can be achieved in
experiments, which opens up a new research field of atomic, molecular and optical
physics. In intense laser fields, the possibility for multi-photon ionization to happen
significantly increases [1], which enables molecules that are stable in laser fields of
usual intensity to get ionized and even fragmented. By focusing the intense laser light,
it is possible to approach the intensity of the order of 2petawatt/cm , which is
comparable to the magnitude of the electric field inside the molecule. With the
irradiation of such strong laser light, molecules are known to dissociate into multiply
charged ions which subsequently experience bond fission processes generating
charged ion fragments with high kinetic energy. This phenomenon is called Coulomb
explosion, which has become a fascinating research target in recent years. By
studying the ion yield, the kinetic energy release and other physical quantities of the
ion fragments, it is helpful to reveal the dynamics of Coulomb explosion and to obtain
information about the parent ions.
2. Experiment
Fig 1. The diagram on the top shows the simple structure and the working principle of the TOF mass
spectrometer. Down left is a photo of the inside of the chamber with the incident direction of laser and
ejection of gas indicated with arrows. Down right is a photo of the three extraction plates.
The experimental setup is indicated in Fig 1. Femtosecond laser pulses (800 nm)
are generated by a laser source at a repetition rate of 1 kHz. After the amplifiers and
pulse compressor, the pulse energy has reached 2.53 mJ/pulse, with the temporal
width of 45 fs measured by Single Shot Autocorrelator. The light beam is focused by a
lens into the chamber, where it meets the molecular effusive beam of sample gas
perpendicularly right between the first two of the three parallel extraction plates of the
time of flight mass spectrometer which accelerates the ion fragments produced in the
various processes of photoionization. The focus area of the laser light as measured to
be -5 22.24×10 cm by CCD camera by assumption of a Gaussian spatial profile, with
the field intensity estimated to be15 22.53×10 W/cm .
Ions with different mass-to-charge ratios acquire the same amount of kinetic
energy but different velocities inversely proportional to the square root of
mass-to-charge ratios in the electric field, with which they fly through a vacuum flight
tube and get separated in space, which will result in a difference of their arrival times
at the micro-channel-plate (MCP) detector placed at the end of the tube. For ion
fragments generated in Coulomb explosion, they get additional initial momentum of
opposite directions compared to other ions, and their peaks are shifted to opposite
ways from the central peak which is formed by ions of the same mass-to-charge ratio
experiencing no Coulomb explosion. Mass spectra are obtained by mass assignment
of the temporal signals recorded and averaged over 3000 sweeps by a digital oscillator
with a sampling rate of 1kHz.
It is crucial to achieve a relatively high mass resolution in order to resolve the
signals of the same ion fragments generated from different multiple charged ions of
one parent molecule. Conventionally, the mass resolution is defined as /m m ,
where m is the smallest deviation between two peak at which they can be
distinguished on the mass spectrum. In our case, the mass resolution can be more
directly signified byshiftm , the deviation of the Coulomb explosion peaks from the
central peak, or from the central mass number if there is no central peak in between. A
larger shiftm for the same kind of ion fragments indicates a higher mass resolution. To
improve the value of shiftm , extension of the original flight tube is conducted by
increasing the length of flight from 502.6 mm to 1477.6 mm as shown in Fig 2, and
tests are run with air as the sample gas. By comparing specifically the peaks of
nitrogen ions on the mass spectra both before and after extension, it is clear that the
mass resolution is highly increased, since shiftm is almost doubled, which is reflected
clearly as the large separation between the side peaks from the central peak in Fig 3.
After reaching this high resolution, experiments on photoionization of acetylene,
dimethylacetylene and methylacetylene are run respectively. A variable neutral
density filter is placed before the pulse compressor to change the intensity of the laser
beam and spectra under different intensities are recorded for each gas sample. For
MRMI, a zero-order half-wave plate is introduced to rotate the polarization of the
laser beam with respect to the direction of the detection axis. The half-wave plate is
rotated manually at an interval of 3 , which corresponds to a 6 interval for the
polarization angle. Thirty high-resolution time-of-flight spectra covering 180 rotation
of polarization are taken for each gas sample to construct the MRMI. The signals for
ion fragments undergone Coulomb explosion are converted to a momentum
distribution to study the fragment-ion pathways. Throughout the experiment, the
pressure in the main chamber is kept sufficiently low to avoid the space charge effect
[10].
Fig 2. On the left is a photo of the flight tube before extension with a flight length of 502.6mm, and on
the right a photo after extension with a flight length of 1477.6mm.
Fig 3. The +N peaks on TOF mass spectra measured before and after extension of the flight tube. It is
clearly seen that the side peaks are much better separated from the central peak after extension, with
the value of mshift almost doubled, which indicates a much better mass resolution.
3. Photoionization of 2 2C H
3.1 Coulomb explosion pathways of 2 2C H
With the light polarization set parallel to the detection axis, all the mass peaks of
ion fragments produced through Coulomb explosion on the TOF mass spectra appear
as pairs on either side of the central mass number as shown in Fig 4a and Fig 4b.
These double peaks show the process that ions are ejected backwards and forwards
during Coulomb explosion, after which the forward fragments are accelerated further
while the backward fragments have to be decelerated first and then switch direction
and get accelerated again, which causes a flight time difference between same type of
ions ejected to different directions, and a arrival time difference compared to ions
without experiencing such a process.
Fig 4a and Fig 4b are mass spectra of acetylene at the field intensity of
14 25.05×10 W/cm and15 22.30×10 W/cm , respectively, with the voltage of MCPs turned
up high in order to get a clear view of the small peaks but the highest ones saturated.
By comparing the signals on these two spectra, it is obvious that 4+C and +
2C both
show up at the higher intensity, which indicates the opening of new explosion
channels or change of the distribution of different explosion pathways in response to
the intensity increase. The reason why the peaks for the mass-to-charge ratio of 3 to
be assigned 4+C instead of
+
3H is that there are only two hydrogen atoms in the
acetylene molecule, and under the irradiation of intense laser light there is a higher
possibility for chemical bonds to be broken than to be formed. Further validation of
this assumption is necessary.
Most of the signals in Figs. 4a-b contain more than one pair of double peaks,
which indicates that for the same type of fragments, they may still differ in their initial
momenta. This is due to the fact that fragments acquire different initial momenta if
they are generated from different parent ions. It is possible to identify Coulomb
explosion pathways with the momentum releases which can be readily obtained from
the TOF mass spectrum. The initial momentum acquired by the ion fragments from
Coulomb explosion can be calculated by 0mv qF t , where t represents half of
the time difference between the double peaks, F the static electric field, and 0v the
initial velocity of the fragment along the flight tube.
Fig 4a.
Fig 4b
Figs 4 a-b. Mass spectra of acetylene at laser field intensities of 14 25.05×10 W/cm and
15 22.30×10 W/cm . The peaks of +
2 2C H are saturated in that the voltage of MCPs are turned up high to
get a clear view of the small peaks. By comparing the signals on these two spectra, it is obvious that
both 4+C and +
2C show up at the higher intensity, which indicates the opening of new explosion
channels or change of the respective yields of different explosion pathways in response to the intensity
increase.
Fig 5a
Fig 5b
Fig 5c
Fig 5d
Fig 5e
Fig 5f
Figs 5 a-f. The experimental time of flight spectra of +
2C , +C , 2+C , 3+C , 4+C and +H in Fig 4b
plotted over momentum of the ion fragments (red lines). The vertical axis represents the signal intensity
of the fragment ions detected by the MCP. The result of the Gaussian fit for each component is shown
by the blue dotted lines.
Table 1. Released momenta (u ms-1
) of +C , 2+C , 3+C , 4+C , +H and +
2C defined as the center of
the Gaussian profiles in the unit of 1amu m s . Close momenta between different species of
fragments are matched by different colors.
Table 2. Possible Coulomb explosion pathways found by matching of initial momenta.
In Figs 5 a-f, the doublet peaks identified for the fragments, +
2C , +C ,
2+C , 3+C ,
4+C and +H in Fig 4b are expanded and rescaled with respect to the Coulomb
explosion momentum release. As shown in Fig 5a, single pair of peaks is assumed for
+
2C , yet tails on the right side of both of the peaks seem quite puzzling. In most cases,
tails extending towards the higher momentum direction are considered evidence of C+
and C from dissociation of C2+
in the accelerating region between the parallel
electrodes of the TOF spectrometer. However, these two tails are extending towards
lower and higher momentum respectively, which requires further research to explain.
For +C , two pairs of peaks stand out with two pairs of tails extending towards the
central peak and the outer region respectively, which signifies two other possible
components. For 2+C , the wide shoulders on outer side of the higher peaks evidently
indicate the existence of a third component. There is also a tiny tail in adjacent to the
two shoulders, which might be another considerably week component. Similarly,
three pairs of doublet peaks are easily identified for 3+C , two for
4+C , and three for
+H with two inner shoulders of the high peaks considered as one additional
component.
By assuming a Gaussian profile in the momentum distribution of the ion
fragments, the least-squares fitting to the peaks are conducted, and absolute values of
the initial momentum defined as the center of the Gaussian profiles are determined
with high precision which are shown in the form in Table 1 in the unit of1amu m s .
The only exception is the fitting for the two inner tails close to the central peak in Fig
5b, where the broad and weak peak structure makes the fitting extremely difficult.
Here, a constraint of the values of the profile centers is imposed so that the
momentum for this component is close to the momentum of the lowest peaks in the
spectrum of +H , so that they might be matched as fragments from the same parent
ions. Also, the precision for the momenta of the shoulder components are relatively
low compared to that of the peaks that are evidently separated from others. In Table
1, ion fragments with similar momenta are assigned by different colors, with which
four explosion pathways are identified and also indicated by arrows in Fig 5 a-f. All
possible pathways are listed in Table 2, but please note that further investigation is
needed to verify them.
+
2C fragments can only be originated from the Coulomb explosion with one or
two +H ions, since there are only two carbon atoms in one molecule. The
momentum of H for the pathway 2+ + +
2 2C H C +H should be around 198.19
1amu m s , but no hint of peaks around this momentum in the momentum distribution
of +H . For the pathway of 3+ + + +
2 2 2C H C +H +H , the signal of 99.09 1amu m s +H
fragments may be submerged by the high peaks at around 107.17 and 85.15 1amu m s ,
since the signal intensity of +
2C is considerably small compared to these two species
of +H fragments, which leaves this pathway undetermined for now.
The 2+C peaks with the lowest momentum apparently match the
+H peaks with
almost the same value of momentum, with which the pathway of 3+ 2+ +(CH) C + H
can be identified, but the existence of this pathway is not perfectly secured due a
reason to be discussed in the next paragraph. The highest peaks of 2+C may originate
from the pathway of 4+ 2+ 2+
2C C + C , since this pathway produces more 2+C
fragments for each time of explosion, but further proof is required to support this
conjecture. For the last pair of peaks with the momentum of 170.301amu m s , no
signals at such momentum are found in the momentum distributions of the other
fragments, which disproves the matching with one single ion. Considering three-body
explosion, it is reasonable to guess a pathway as 4+ + 2+ +
2(C H) H +C +C with the
relationship among the fragment momenta as shown in Table 3.
Table 3. Three possible combinations of momenta of the ion fragments in the pathway of
4+ + 2+ +
2(C H) H +C +C .
For 3+C , the lowest inner peaks can be assigned to
+H with similar momentum as
fragment from the same pathway 4+ 3+ +(CH) C + H . However, if such a pathway
exits, it will induce an interesting result that the momentum of the ion fragments from
pathway 3+ 2+ +CH C + H (about 108
1amu m s ) is higher than that from
4+ 3+ +CH C + H (85.15 1amu m s ), while normally a higher momentum release is
assumed for higher repulsive force between the fragments, which in this case 3+C
and +H . Accordingly, either of these two pathways can be wrong. The peaks for
2+C can be explained by three-body explosions instead as described in last paragraph,
but no suitable three body explosion is found that can match the momenta of the
fragments. The 3+C peaks with the momentum of 115.98
1amu m s are almost
certainly matched to 4+C with the momentum of 114.51
1amu m s , which means the
3+C with momentum of 135.0651amu m s can only be produced by three-body
process since there can be no other two-body pathway for 3+C fragments releasing
larger momentum than 7+ 3+ 4+
2C C + C .
Now in Fig 5b, if the two highest pairs of peaks are matched with 3+C and
2+C
for either two or three body explosion, the lowest peaks with smaller momentum can
only be matched to carbon ion with a smaller charge number which is no other but
+C . Of course, they may also originate from the three body pathways mentioned
previously.
The last pair of doublet peaks of carbon ions, the 101.3 1amu m s 4+C
fragments has no other fragments to match it for a two body explosion, but a three
body explosion pathway 6+ 4+ + +
2(C H) C +C + H is conceivable with the momentum
of +H measured as 46.9751amu m s and
+C calculated as 54.3291amu m s
according to the conservation of momentum along the flight axis. It is possible for
+C with such a value of momentum to exist, since the shape of the peaks at around
this value is severely spoiled by the noisy fluctuation and the position of the profile
center is quite ambiguous.
Only the mass spectra of +H and +C contain a central peak in the middle of the
double peaks, which represent ions with no initial momentum from Coulomb
Explosion. It is important to note that although the central peak ion fragments are
mostly generated by direct dissociation from the molecule, it is also highly possible
that such signals contain components of pathways where uncharged atoms are
produced from the explosion, leaving the other charged fragments with minute initial
momentum, such as +CH C+H and CH C +H .
3.2 Laser intensity dependence of ion yield
Fig 7. Lase field intensity dependence of 2+C ion yield based on ten TOF spectra measured at different
intensities. The ion yield takes into account the left side peaks of all the components of 2+C . It is clear
that the ion yield is roughly linear to the laser field intensity.
With TOF spectra measured at ten different laser-field intensities, the intensity
dependence of the ion yield of 2+C fragments is plotted as shown in Fig 7. Due to
limited time, the calculation of ion yield is only performed on the sum of the signals
of 2+C on the left side of the doublet, so this result shows a synthetic response of
several pathways to the change of laser field intensity. It can be seen in Fig 7 that the
ion yield is roughly linear to intensity, yet the linearity is bad when values are
converted to log scale.
There are two factors necessary in understand the intensity dependence. First, the
Coulomb explosion in this experiment should be a multi-photon process, which has
been explained the first part of this report. The numbers of photons required to reach
the bond energy of several types of bonds are listed in Fig 8. Since different bonds are
broken in different pathways, it will be helpful to study the intensity dependence for
each individual pathway as well. The other factor, the volume effect, makes the
situation a lot harder to be analyzed. Generation of different types of ions may tend to
dominate areas at different radii to the center of the focused laser beam (indicated in
Fig 9), or in other words the laser intensity changes gradually around the focus point,
which will result in a spatial distribution of the ion yields of the ion fragments. The
difficulty lies in how the radii of different ions change with respect to the laser field
intensity, or how the spatial distribution changes in response to the change of intensity.
Thus, a lot more research is needed in the future to understand the laser field intensity
dependence.
Fig 8. Bond energies of C- H , C-C , C = C , and C C , and numbers of photons at 800nm
required to break each of these bonds.
Fig 9. A simple illustration of how volume effect takes place. Around the focus point of the laser beam,
the laser field intensity changes gradually, which will result in a spatial distribution of the ion yields of
different ion fragments.
Fig 10. Ion yield of 2+C as a function of polarization of the laser light. The red crossings show the total
ion yields of the left peaks of 2+C at thirty different laser polarizations. The profile exhibits
symmetric structure as expected. The signals measured at the polarization angles from 60 to 132
are considered as zero, which makes it difficult to fit the curve with an appropriate function.
3.3 Polarization dependence of ion yield
Polarization dependence of the ion yield of 2+C is plotted with thirty TOF
spectra measured by manually rotating the direction of light polarization for a full
180 angle at an interval of 6 . Again, the calculation of ion yield is only performed
on the sum of the signals of 2+C on the left side of the doublet due to limited time.
As shown in Fig 10, the profile exhibits symmetric structure as expected.
Unfortunately, the signals measured at the polarization angle from 60 to 132 are so
small that they are submerged in the background, for which the ion yield can only be
considered as zero, therefore part of the information is lost and it makes it difficult to
fit the curve with an appropriate function. More accurate experiments may help to
determine the profile of polarization dependence. It is also advisable to study the
polarization dependence for individual Coulomb explosion pathways after they are
confirmed, since the dependence profile of each pathway may also be different.
4. Ejection of +
3H in dimethylacetylene and methylacetylene
Fig 11. +
3H signals (indicated by blue arrow) on an expended mass spectrum of dimethylacetylene at
the laser field intensity of 15 22.47×10 W/cm . The peaks are so small compared to +
2H that they are
merely recognizable in the background noise.
Fig 12. +
3H signals (indicated by blue arrow) on expended mass spectrum of methylacetylene at the
laser field intensity of 15 22.38×10 W/cm . Two pairs of peaks are clearly observed whereas only single
doublet peaks are observed in previous studies at a lower laser field intensity, which indicates the
possible existence of a new pathway.
TOF mass spectra of dimethylacetylene and methylacetylene are processed and
expended in Fig 11 and Fig 12 to see the +
3H signals at the laser field intensity of
15 22.47×10 W/cm and 15 22.38×10 W/cm respectively. In previous studies, ejection of
triatomic hydrogen ions is observed for methylacetylene at the laser field intensity of
13 25.0×10 W/ cm , where the signal intensity of the single pair of +
3H peaks is
sufficiently large to be comparable to the peaks of +
2H [8]. However, as shown in Fig
11, although there are more 3CH in dimethylacetylene, the +
3H peaks at m/z = 3 are
so small compared to +
2H at m/z = 2 that they are merely recognizable in the
background noise. One possible explanation for the low yield of +
3H of
dimethylacetylene is that +
3H are generated at excited states due to the high intensity
of the laser that most of them are not stable enough to survive the flight and reach the
detector. For methylacetylene, the peaks at m/z = 3 are a lot larger, but two pairs are
clearly observed in Fig 12, and it is reasonable to guess that the outer pair of them to
be C4+
, since the mass-to-charge ratio of C4+
is also 3, and the momentum release of
C4+
should be higher than that of +
3H . These are no doubt interesting topics to look
into for future research.
5. Summary
In this project, the high resolution of time-of-flight mass spectrometer is achieved
by extending the length of the flight tube, with which photoionization of acetylene,
dimethylacetylene and methylacetylene in an intense laser field (45 fs, 800 nm,
15 210 W/ cm ) is studied by recording spectra at different laser intensities and
different polarizations of the light. By deriving the momentum releases of Coulomb
explosion and analyzing the profile features of the fragment signals, five two-body
pathways and three three-body pathways are found possible for acetylene. A relatively
simple analysis of intensity dependence and polarization dependence of the Coulomb
explosion of acetylene is conducted on the basis of high-resolution TOF mass spectra.
Features of triatomic hydrogen ions ejected by dimethylacetylene and
methylacetylene different from previous studies are observed. A few questions about
the interpretation of the experimental results also rise along the analysis, and they can
make valuable topics for future research.
Acknowledgements
I would like to thank Professor Yamanouchi for providing me this position in his
fabulous lab, and for his enlightening discussions and critical reading of the draft of
this report, and S. Owada San for help and discussions throughout the experiment, and
all the other lab mates for making my stay enjoyable. Thanks also to all the UTRIP
coordinators who worked hard for the success of this program.
References
1. P. B. Corkum, Physical Review Letters, Vol. 71, No. 13 (1993).
2. Michael Guilhaus, Journal of Mass Spectrometry, Vol. 30, 1519-1532 (1995).
3. L. J. Fransinski, K. Codling, P. A. Hatherly, Science 46 (1989) 1029.
4. C. Cornaggia, M. Schmidt, D. Nirmand, J. Phys. B 29 (1996).
5. A. Hishikawa, A. Iwamae, K. Hoshina, M. Kono, K. Yamanouchi, Chemical
Physics 231 (1998) 315-329.
6. A. Hishikawa, A. Iwamae, K. Hoshina, M. Kono, K. Yamanouchi, Chemical
Physics Letters 282 (1998) 283-291.
7. A. Hishikawa, A. Iwamae, K. Yamanouchi, Journal of Chemical Physics, Vol. 111,
No. 19 (1999).
8. T. Okino, A. Watanabe, H. Xu, K. Yamanouchi, Phys. Chem. Chem. Phys,
2012,14, 4230-4235.
9. Y. Furukawa, K. Hoshina, K. Yamanouchi, H. Nakano, Chemical physics Letters
414 (2005) 117-121.
10. C. Cornaggia, J. Lavancier, D. Normand, J. Morellec, H. X. Liu, Phys. Rev. A 42
(1990) 5464.